Genome-Wide Identification and Characterization of the SHI-Related Sequence Gene Family in Rice

Data acquisition

The rice sub-species, maize, sorghum, Brachypodium genomic and proteomic sequences were downloaded from Phytozome (https://phytozome.jgi.doe.gov), Ensemble plants (http://plants.ensembl.org/index.html), and PlantTFDB (http://planttfdb.cbi.pku.edu.cn/) ¹⁵ ; the SRS protein sequences of Arabidopsis thaliana were obtained from the Arabidopsis genome database (TAIR 9.0 release, http://www.arabidopsis.org/) ¹⁶ ; and the OsSRS gene expression data were collected from the SRA database (https://www.ncbi.nlm.nih.gov/).

SRS gene identification and characterization

The hidden Markov model (HHM) file of the SRS domain (PF01542) was downloaded from Pfam protein family database (http://pfam.sanger.ac.uk/). ¹⁷ HMMER 3.0 was used to search the SRS genes from the maize, sorghum, Brachypodium, and Arabidopsis genome database. The default parameters were adopted and the cutoff value was set to 0.01. All candidate genes that may contain DUF247 domain or RING-like zinc finger domain based on HMMER ¹⁸ results were further examined by confirming the existence of SRS core sequences using the PFAM and SMART program (http://smart.embl-heidelberg.de/).^19,20 Each potential gene was checked manually to ensure the conserved 28-peptide region of the predicted SRS domain. The incorrectly predicted genes were manually curated and the redundant sequences were discarded. Six transcripts were finally identified in the rice genome after comprehensive curation.

The other gene features such as protein length, molecular weight, isoelectric point, the possible sub-cellular localizations and trans-membrane area, were predicted using ExPASy (https://www.expasy.org/). ²¹

DNA-binding domain conservation analysis of OsSRS proteins

The complete protein sequences of OsSRS were retrieved by Quick Fasta Extractor or Filter software ²² (https://github.com/CJ-Chen/TBtools). Then, we identified all highly conserved domain for all OsSRS members by multiple sequence alignment analysis using ClustalW (https://www.genome.jp/tools-bin/clustalw) and visualized in JavieW software (http://circinus.ebi.ac.uk:6543/jalview/). ²³

Gene structure of OsSRS proteins

The positional information of the OsSRS genes (intron-exon GTFF/GFF3 files and genome sequence files) was loaded into the Gene Structure Display Server (GSDS v2.0, available online: http://gsds.cbi.pku.edu.cn/) to obtain information on UTR/Intron/CDS structure. ²⁵

Conserved motif detection

The conserved motifs of SRS proteins were detected using Multiple Expectation Maximization for Motif Elucidation (MEME v4.11.2, Available online: http://meme.nbcr.nrt/meme/) ²⁶ with the parameters set as maximum width 5, minimum width 150, and maximum number 10.

Chromosomal localization, detection of gene duplication events, syntenic analysis and estimation of synonymous(Ks) and non-synonymous(Ka) substitution per site and their ratio

According to the gene positional information, we draw a map of the distribution of OsSRS genes throughout the rice genome using Amazing Gene Location software of TBtools. ²² The OsSRS gene pairs resulting from segmental duplication were linked by red lines.

Duplicated gene pairs of OsSRS genes within the japonica rice derived from segmental or tandem duplications were identified in the rice genome based on the method in the Plant Genome Duplication Database. An all-against-all BLASTP comparison with an E-value ⩽ 1 x 10^-20 was performed for syntenic clustering determined by MCscanX (E-value ⩽ 1 x 10^-20). ²⁷ Pairs from segmental duplications were used to estimate Ka, Ks, and their ratio. The coding sequences from segmentally duplicated OsSRS genes were aligned by ClustalW and trimmed by Gblocks. The software of TBtools was then used to compute Ka, Ks values for each pair following the YN model. ²⁸ The result is that if Ka/Ks > 1, it is positive selection pressure; if Ka/Ks = 1, it is neutral selection or natural selection pressure; if f Ka/Ks < 1, it is purification selection pressure. ²⁹

Plant growth conditions, RNA sampling and qRT-PCR analysis of OsSRS genes

Indica rice IR26 seeds were grown at 28℃ and 12/12 hours light/dark period in the environmentally controlled greenhouse. Different organs of rice seedlings were collected after 20 days at the tillering stage. All the samples were kept in liquid nitrogen, and stored at -80℃ for further use. High-quality RNA was extracted using TRI REAGENT RT reagent (Molecular Research Center, Inc.) following the manufacturer’s instructions (http://www.mrcgen.com/tri-rt.ht).

The real-time PCR (qPCR) assay was completed to detect the relative expression of SRS genes in different rice organs as previously described. ³⁰ Original expression data for OsSRS genes under different treatments including GA, paclobutrazol (PB), salt (PRJEB4672) and drought stress (PRJNA272723) were retrieved from NCBI SRA (Available online: https://www.ncbi.nlm.nih.gov/). All the data were analyzed using RNA-seq analysis process, including TopHat, cufflinks, cuffmerge, cuffcompare, cuffquant, and using cuffnorm to perform normalization and log transformation. Heat-map was produced by Amazing Heat Map software of serials TBtools ²² with average linkage hierarchical clustering.

Identification and characterization of OsSRS genes in O. sativa subsp. japonica

To extensively understand the SRS gene family in japonica rice, we searched MSU rice databases with Pfam ¹⁷ Number PF05142 and verified by SMART software.^19,20 Eventually, only six candidate DUF702 genes in japonica rice were identified; these genes coincidently were all later found to contain the SRS gene.

In addition, we systematically identified the basic protein properties of SRS candidates, including domain position, isoelectric point, molecular weight, cellular localization, and transmembrane region (Table 1). The length of the SRS proteins ranged from 303 aa (LOC_Os08g43410) to 360 aa (LOC_Os06g49830). Correspondingly, the molecular weights ranged from 30.69 kDa (LOC_Os08g43410) to 34.95 kDa (LOC_Os06g49830), and the predicted isoelectric points varied from 7.27 (LOC_Os01g71490) to 8.97 (LOC_Os09g36160). Four proteins encoded by SRS members were predicted to localize on the cell membrane and in the nucleus, while the other two were only predicted to localize in the nucleus. The cell membrane localization indicated that these proteins would have contact with external signals outside of the cells in the environment; however, there were no trans-membrane regions detected among these rice SRS proteins. The locations of the DUF702 and SRS domains were identified using SMART and positions of domains in similar SRS proteins, such as LOC_Os01g72490.1 and LOC_Os01g72490.2 (the model for alternative splicing of the SRS domain). LOC_Os08g43410 and LOC_Os09g36160 (phylogenetically closely related), encoded by the same or closely related genes, were similar, but quite different from the proteins encoded by phylogenetically alienated related genes, such as LOC_Os05g32070 and LOC_Os06g49830.

OPEN IN VIEWER

Table 1.

The information of SRS proteins in rice.

Gene ID	Amino acid length	Domain position	pI	Mw	Cellular localization	Trans-membrane analysis
LOC_Os01g72490.1	341	114-269	7.27	34.77	Cell membrane. Nucleus	No tans-membrane area
LOC_Os01g72490.2	341	114-269	7.27	34.77	Cell membrane. Nucleus	No tans-membrane area
LOC_Os05g32070.1	331	108-269	8.76	33.18	Cell membrane. Nucleus	No tans-membrane area
LOC_Os06g49830.1	360	131-286	8.68	34.95	Cell membrane. Nucleus	No tans-membrane area
LOC_Os08g43410.1	303	91-229	8.67	30.69	Nucleus	No tans-membrane area
LOC_Os09g36160.1	316	95-242	8.97	32.02	Nucleus	No tans-membrane area

Abbreviations: Mw, molecular weight; pI, isoelectric point; SRS, SHI related sequence.

Multiple sequence alignment and phylogenetic analysis

SRS TFs usually have a highly conserved 138–161-aa region in the N-terminus, which contains a putative RING-like zinc finger unique to SRS proteins. The zinc finger protein has a C-X2 C-X7 C-X4 C-X2 C-X7 C arrangement (where X represents variable amino acids), and the conserved motif is essential for its regulatory function. In this study, the OsSRS family’s conserved domains were analyzed. Multiple alignments of the OsSRS protein sequences against SRS-binding domains revealed that they were highly conserved. In particular, a highly conserved domain structure C-QD-C-GNQAKKD-C-GHQR-C-RT-C-CKSRGFD-C was found in the OsSRS family (100% alignment) in a 28-aa region having the structure C-X2 C-X7 C-X4 C-X2 C-X7 C (Figure 1).

OPEN IN VIEWER

Figure 1.

Multiple alignments of six OsSRS protein sequences against the putative zinc finger domain. The identical zinc finger domains of eight cysteine residues are shown on the top, and the dashed rectangle indicates the DUF702 domain area.

In addition, the domain positions in the OsSRS protein family were correlated with the members’ positions in the phylogenetic tree. The DUF702 domain contains two other conserved domains in the rice SRS proteins. Based on the phylogenetic tree, the domain structures in members of the same sub-group were nearly identical (Figure 2A).

OPEN IN VIEWER

Figure 2.

The gene features of OsSRS genes. (A) The domain structure and phylogenetic tree of OsSRS protein family. (B) OsSRS gene structures and phylogenetic tree. UTR, CDS, and introns were depicted by filled blue boxes, yellow boxes, and single black lines. The phylogenetic tree was constructed by MEGAX with bootstrap 1000 times. (C) Motif 1, corresponding to C-X2 C-X7 C-X4 C-X2 C-X7 C zinc finger conserve structure, the detailed motif structure was shown in Supplementary Table S4. CDS indicates coding sequence; UTR, un-translated region.

To explore the phylogenetic relationships of OsSRS proteins, we carried out a phylogenetic analysis of SRS proteins from rice sub-species. A phylogenetic tree was constructed to include three cultivar rice species, six wild rice species, one BB-genome rice species, and one FF-genome rice species (Figure 3, Supplementary Table S1). For each gene, we chose all the proteins formed by alternative splicing. All the rice species contained five SRS genes, except O. barthii, O. nivara, O. longistaminata, and O. brachyantha. The phylogenetic tree revealed that all the SRS proteins in the various rice sup-species clustered into five major groups (A, B, C, D and E), and they were considered to be distinct based on bootstrap values greater than 80.

OPEN IN VIEWER

Figure 3.

Phylogenetic relationships among AA genome rice with three cultivars as O. sativa japonica, O. sativa indica, O. glaberria, six wild rice as O. nivara, O. rufipogon, O. barthii, O. longistaminata, O. meridonalis, and O. glumaepatula. one BB genome rice O. punctate and one FF genome rice O. brachyantha. The neighbor-joining tree was created by MEGAX with bootstrap value setting at 1000. All the SRS proteins were divided into five major groups. Different color represents different groups divided based on the bootstrap values: the small red star represents values between 50% and 80% and the big red star represents bootstrap values between 80% and 100%. Detailed information was listed in Supplementary Table S1. SRS indicates SHI related sequence.

Group A contained 14 SRS proteins, groups B, C and E contained 10 protein members each, and group D had 7 members. Interestingly, the SRS proteins in the rice species were distributed evenly. In group A, SRS proteins from O. glaberrima and O. barthii clustered together with indica, derived from O. nivara, had a closer relationship with japonica, derived from O. rufipogon. The other three SRS proteins from the AA-genome rice species O. logistaminata, O. glumaepatula, and O. meridionalis were further apart. All the SRS proteins from AA-genome rice species clustered together, but SRS proteins of the BB- and FF-genome rice species O. punctata and O. brachyantha, respectively, clustered as out groups from AA-genome rice species. The results were similar for the B, C, D and E groups.

Furthermore, we constructed a phylogenetic tree of SRS proteins from one dicot plant A. thaliana and other four monocot plants rice, maize, sorghum, and Brachypodium (Figure 4A). The SRS proteins were divided into 11 major groups, from group A to K, supported by bootstrap values greater than 80%. Group A was the largest group, containing three, four, and two SRS proteins in rice, Brachypodium, sorghum, and maize, respectively. Groups B and C each contained seven SRS members with three maize SRS proteins in both clades. Group B also contained two, one and one members from rice, sorghum, and Brachypodium, respectively, while group C also contained one, two and one members from rice, sorghum, and Brachypodium, respectively. Group D contained three maize members and one member each from Brachypodium and sorghum. The groups from E to K were all A. thaliana members and contained no monocot protein members, except one Brachypodium member in group I. The SRS proteins in monocot and dicot plants diverged, which could result from species-/lineage-specific gene gain or loss events. We further checked the gene ontology (GO) annotations of these japonica OsSRS and Arabidopsis proteins. Some groups had specific annotations, such as group E, which was involved in “stigma development” and “xylem and phloem pattern formation.” In addition, group I contained only one A. thaliana (AT1G32730) member that annotated as “located in membrane.” Groups F and G contained members that were annotated as playing roles in “root development” and “gynoecium development” compared with the other groups, respectively. AT1G19790 was involved in “sequence-specific DNA binding” in group G (Supplementary Table S2). All these specific annotations in different groups from E to K imply possible divergent functions of the SRS genes in these groups. The phylogenetic tree showed that the SRS genes in monocot plants were duplicated several times before the divergence of these four species, and were highly conserved in rice, Brachypodium, maize, and sorghum. The SRS proteins from the monocot plants may be more closely related to each other evolutionarily to those from the dicot plant A. thaliana. In addition, the SRS gene numbers and proteins were also analyzed in Figure 4B. There were 11 gene members and 14 protein members in Arabidopsis; however, the genes members were only one less than protein numbers in monocot plants we had detected in this study. For example, in rice, there were only 5 SRS genes but 6 protein members, and in maize, 10 SRS proteins were found but 9 SRS genes were detected. Interestingly, the SRS members in dicot plants such as Arabidopsis were more than in monocot plants such as rice, maize, which may be because the annotation of the monocot genome was not too perfect.

OPEN IN VIEWER

Figure 4.

Phylogenetic relationships among four monocot plants including rice, Brachypodium, maize and sorghum, and one dicot plant Arabidopsis thaliana. The neighbor-joining tree was created by MEGAX with bootstrap value setting at 1000. The SRS proteins from monocot plants were conserved and divided into four clades (clades A to D) and the rest of the SRS proteins were mainly from A.thaliana, except for one member from Brachypodium. The phylogenetic tree indicates the SRS protein function divergence between monocot and dicot plants. Different color represents different groups divided based on the bootstrap values: the small red star represents bootstrap values between 50% and 80% and the big red star represents bootstrap values between 80% and 100%. Detailed information was listed in the Supplementary Table S3. SRS indicates SHI-related sequence.

Gene structure and chromosome location analysis

Gene organization is important in the evolution of gene families, and the gene structure is the backbone. Therefore, we compared the introns and exons of OsSRS genes by aligning the genomic DNAs (Figure 2B). The results were consistent with the phylogenetic tree in that the intron/exon structure patterns were similar within the same group. For example, based on the gene structures, LOC_Os08g43410 and LOC_Os09 g 36160 contained 1 intron, 2 CDSs and 2 UTRs, which was different from the structures of the other groups. LOC_Os06g48830 and LOC_Os05g32070 had different structures, while LOC_Os01g72490.1 and LOC_Os01g72490.2 were two spliced forms of LOC_Os01g72490.

Our classification was also verified by the conserved motif analysis. All of the SRS proteins from various rice species were loaded into the MEME analysis tool to identify the conserved motifs. Information on all 10 motifs observed with statistically significant E-values below 5.10E-189 can be found in Supplementary Table S4 and Figure 5. The motifs identified by MEME were between 9 and 52 aa in length. One of the motifs (motif 1), corresponding to the C-X2 C-X7 C-X4 C-X2 C-X7 C zinc finger conserved structure, was found in all the SRS proteins in rice (Figures 2 C and 5). Motif 1 was also common in all the SRS proteins from various sub-species, except most of the SRS proteins from the BB- and FF-genome species O. punctata and O. brachyantha, respectively, making it a highly conserved motif of the SRS family. In the phylogenetic tree of SRS proteins from rice subspecies, groups C and D shared a close relationship with the same motif types, but their motif arrangements were different. For example, group C had a motif arrangement “10-3-8-1-5-2-8-4,” while group D had a motif arrangement “10-5-8-1-3-2-8-4.” However, some members did not comply with these arrangements, maybe because of the different genome types, such as OB09G24020.1 and OPUNC09G15790.1. The SRS proteins from group A contained the largest number of motifs and, specifically, contained both motifs 9 and 6. Nearly three of the SRS proteins that did not contain motif 1 were from group E. In group B, members had complicated motif patterns compared with the other groups, and the motif arrangement “3-2-8-4-7” was specific to this group.

OPEN IN VIEWER

Figure 5.

Conserved motifs of SRS proteins in rice subspecies. The distribution of 10 conserved motifs in SRS proteins. Different colors represent the different groups (A to E) based on their protein sequences constructed a phylogenetic tree. SRS indicates SHI related sequence.

We have also investigated the chromosomal distribution of OsSRS genes which were plotted using gene location software (https://github.com/CJ-Chen/TBtools) (Figure 6). The results indicated that the OsSRS genes were distributed evenly in five chromosomes that were chromosome 1,5,6,8,9.

OPEN IN VIEWER

Figure 6.

Chromosomal locations and gene duplication events of OsSRS. The scale on the left is basepair (bp). The SRS gene pairs of segmental duplications are linked by red lines. Detailed information of duplicated pairs is described in Supplementary Table S5. SRS indicates SHI-related sequence.

Gene duplications and synteny analysis of SRS genes

Gene duplication plays an important role during the process of plant evolution. Gene family expansion and genomic evolutionary mechanisms mainly depend on gene duplication events. In this study, gene duplication events were identified in the SRS gene family. The chromosomal distribution of OsSRS genes was plotted using gene location software (https://github.com/CJ-Chen/TBtools) (Figure 6). The duplication events were also examined, but no tandem duplications were detected in SRS genes from japonica. As shown in Figure 6, two segmental duplications occurred between two pairs of OsSRS genes. These two segmentally duplicated pairwise genes were evolutionary close together in the phylogenetic tree and had very high sequence similarities (Figure 2B), which indicated that the SRS gene expansion in rice mainly resulted from segmental duplications, while tandem duplications played no role. Gene duplicates are usually closely adjacent to each other in the same chromosome, a phenomenon that is speculated to result from unequal crossing over.

To further understand the evolutionary selection on this gene family, we calculated the non-synonymous substitution rate (Ka), synonymous substitution rate (Ks), and Ka/Ks for these two pairs of duplicated genes (Supplementary Table S5). These two pairwise genes’ Ka/Ks were less than one, which implied that the SRS genes in rice tended to be subjected to purifying selection. This may play an important role in the possible functional divergence among OsSRS genes.

To further infer the phylogenetic mechanisms of rice SRS family, we constructed three comparative syntenic maps of rice associated with three representative species including one dicot plant Arabidopsis.thaliana, two monocot plants Brachypodium and sorghum (Figure 7). Notably, all the rice SRS members showed syntenic relationship with those in Brachypodium and sorghum, but only one for dicot plant Arabidopsis and rice (Supplementary Table S6). The numbers of orthologous pairs between the other three species (Arabidopsis, Brachypodium, and sorghum) were 1, 6, and 6, which were LOC_Os01g72490.1-AT3G54430.1, LOC_Os01g72490.1-Bradi2g61120.1, LOC_Os01g72490.1-Bradi2g26823.1, LOC_Os05g32070.1-Bradi2g61120.1, LOC_Os06g49830.1-Bradi1g34990.1, LOC_Os09g36160.1-Bradi3g42197.1, LOC_Os09g36160.1-Bradi4g36030.1, LOC_Os01g72940.1-Sobic.003G429800.1, LOC_Os01g72940.1-Sobic.009G122600.1, LOC_Os05g32070.1- Sobic.009G122600.1, LOC_Os06g49830.1-Sobic.010G261700.1, LOC_Os09g36160.1-Sobic.002G274500.1, LOC_Os09g36160.1-Sobic.007G179000.1. All the rice SRS members were found to be associated with 6 gene pairs such as related to the phylogenetic relationship between rice and other 2 monocot species. Interestingly, the other 4 rice SRS members were not identified as collinear gene pairs between rice and Arabidopsis. Combined with the result of phylogenetic analysis, it may indicate that these orthologous pairs formed after the divergence of dicotyledonous and monocotyledonous plants. And the common identified between rice and 2 other monocot plants also indicated that these orthologous pairs may already exist before the ancestral divergence.

OPEN IN VIEWER

Figure 7.

Syntenic analysis of SRS genes between rice and other three representative plant species. The gray lines indicate the collinearity blocks within japonica rice and other plant genomes. (A) Collinearity between rice and Arabidopsis. (B) Collinearity between rice and Brachypodium. (C) Collinearity between rice and sorghum. The detected gene pairs were also presented. at indicates Arabidopsis chromosome; Bd, Brachypodium chromosome; os, rice chromosome; sb, Sorghum chromosome; SRS, SHI-related sequence.

To better understand the evolutionary constraints acting on the SRS gene family, the Ka/Ks ratios of the SRS gene pairs were calculated. All the orthologous pairs had Ka/Ks < 1 except the Ka/Ks ratio between rice and Arabidopsis with unknown reasons, suggesting that the SRS family might have experienced strong purifying selective pressure during evolution.

Expression patterns of OsSRS genes under different abiotic stresses and hormonal treatments

To examine the possible divergent functions of OsSRS genes, we investigated their expression levels in different organs including root, shoot, leaf, panicle, seed, anther, and pistil with qRT-PCR. The results showed that LOC_Os01g72480 expressed higher than other members in the organisms of root, shoot, leaf, and panicle. All the SRS genes could be detected in the above-mentioned organisms except leaf and seed. In leaf, only three genes were detected, and LOC_Os01g72480 expressed highly, otherwise, LOC_Os05g32070 and LOC_Os08g43410 expressed extreme lowly. Moreover, LOC_Os08g43410 expressed extreme highly in seed, but the other three genes (LOC_Os05g32070, LOC_Os06g49830 and LOC_Os09g36160) were not detected in seed (Figure 8A to G).

OPEN IN VIEWER

Figure 8.

Expression patterns of OsSRS genes in different organs and under different treatments. (A) to (G), Relative expression of SRS genes in different organs in rice including root, shoot, leaf, panicle, seed, anther and pistil; (H) The heatmap was generated by Amazing Simple HeatMap of TBtools software using the expression data of OsSRS genes and normalized log₂ transformed values were used with hierarchical clustering represented by the color scale (1-6). Green indicates as low expression and the red indicates high expression. The samples were leaves of IR29 rice seedlings under GA, PB, salt (different times), and drought treatments. Detailed information of expression data is described in Supplementary Tables S6 and S7. GA indicates gibberellin; PB, paclobutrazol; SRS, SHI related sequence.

In addition, we also checked the SRS gene expression levels under various abiotic stress conditions and during hormonal treatments in the indica rice IR26 using the public transcriptome data from the NCBI SRA database (Supplementary Tables S7 and S8). We used a heat-map to visualize the global transcriptome profile of these OsSRS genes, and the results present the diverse responses to various treatments, including GA, PB, salt, and drought (Figure 8H).

In the GA- and PB-treatment experiments (gene expression data collected from leaves of rice seedlings), compared with the negative control mock, one gene, LOC_Os01g72490, was induced by GA but depressed by PB, while the other members were slightly depressed by GA, and the pattern was opposite for PB. We further checked the GO annotations of their Arabidopsis orthologs and found that they were annotated as “auxin responsive,” “root and pollen development,” “anther development” and “multicellular organism development,” which may imply possible functional roles for these genes.

We also further analyzed the expression patterns of OsSRS genes in 300 mM NaCl- and drought-stressed seedlings at the following different times: 1, 2, 5, 10 and 24 hours after salt stress within a day. The expression data were calculated between abiotic treatments and controls. Some of them showed significant changes under salt-stress conditions. LOC_Os01g72490 appeared to be down-regulated 2 and 24 hours after treatments but up-regulated 5 hours after the salt treatment, which indicates that its expression first increased, then decreased and increased again. LOC_Os06g49380’s expression was depressed by the salt treatment compared with the mock. The expression levels of other members did not change under salt-stress conditions. However, under drought-stress conditions, one gene (LOC_Os01g72490) was repressed, two genes (LOC_Os08g43410 and LOC_Os09g36160) were induced, one gene (LOC_Os05g32070) was insensitive and one gene (LOC_Os06g49830) was not detected. Notably, gene expression patterns consistent with the gene positions in the phylogenetic tree.